A Comparative Study on Plant Disease Detection Using Deep Learning Techniques

Agriculture is becoming a significant business in Bangladesh due to the country's expanding population and rising food demands. Consequently, it is imperative to raise agricultural productivity. Among the important factors influencing lower agricultural yields are diseases caused by bacteria, fungi, and viruses. Techniques for detecting plant diseases can be used to control and prevent this. Since deep learning techniques primarily use plant data and provide very effective means of identifying plant ailments, they will be used for the diagnosis of plant diseases. Using deep learning and machine learning based on artificial intelligence (AI), the method offers a thorough assessment of all the methods used in plant disease detection. For our study, we have considered Convolutional Neural Network (CNN), MobileNetV2 and Inception V3techniques. Following preprocessing, the models receive 80% of the datasets for training purposes and the remaining 20% for testing. After the disease has been detected, performance analysis and model comparison are carried out. It is found that MobileNetV2 outperforms the other applied approaches, yielding an accuracy of 98.43 %.